arXiv subjects

Akansha Singh

Publications and source records attributed to Akansha Singh.

6 recordsLinked to original sources

arXiv2025

Generative AI in Depth: A Survey of Recent Advances, Model Variants, and Real-World Applications

In recent years, deep learning based generative models, particularly Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Diffusion Models (DMs), have been instrumental in in generating diverse, high-quality content across various domains, such as image and video synthesis. This capability has led to widespread adoption of these models and has captured strong public interest. As they continue to advance at a rapid pace, the growing volume of research, expanding application areas, and unresolved technical challenges make it increasingly difficult to stay current. To address this need, this survey introduces a comprehensive taxonomy that organizes the literature and provides a cohesive framework for understanding the development of GANs, VAEs, and DMs, including their many variants and combined approaches. We highlight key innovations that have improved the quality, diversity, and controllability of generated outputs, reflecting the expanding potential of generative artificial intelligence. In addition to summarizing technical progress, we examine rising ethical concerns, including the risks of misuse and the broader societal impact of synthetic media. Finally, we outline persistent challenges and propose future research directions, offering a structured and forward looking perspective for researchers in this fast evolving field.

cs.CV↗

arXiv2024

Piecewise Padé-Chebyshev Reconstruction of Bivariate Piecewise Smooth Functions

We extend the idea of approximating piecewise smooth univariate functions using rational approximation introduced in \cite{aka_bas-19a} to two-dimensional space. This article aims to implement the novel piecewise Maehly based Padé-Chebyshev approximation \cite{mae_60a}. We first develop a method referred to as PiPC to approximate univariate piecewise smooth functions and then extend the same to a two-dimensional space, leading to a bivariate piecewise Padé-Chebyshev approximation (Pi2DPC) for approximating piecewise smooth functions in two-dimension. We study the utility of the proposed techniques in minimizing the Gibbs phenomenon while approximating piecewise smooth functions. The chief advantage of these methods lies in their non-dependence on any apriori knowledge of the locations and types of singularities (if any) present in the original function. Finally, we supplement our methods with numerical results to validate their effectiveness in diminishing the Gibbs phenomenon to negligible levels.

math.NA↗

arXiv2023

ETMA: Efficient Transformer Based Multilevel Attention framework for Multimodal Fake News Detection

In this new digital era, social media has created a severe impact on the lives of people. In recent times, fake news content on social media has become one of the major challenging problems for society. The dissemination of fabricated and false news articles includes multimodal data in the form of text and images. The previous methods have mainly focused on unimodal analysis. Moreover, for multimodal analysis, researchers fail to keep the unique characteristics corresponding to each modality. This paper aims to overcome these limitations by proposing an Efficient Transformer based Multilevel Attention (ETMA) framework for multimodal fake news detection, which comprises the following components: visual attention-based encoder, textual attention-based encoder, and joint attention-based learning. Each component utilizes the different forms of attention mechanism and uniquely deals with multimodal data to detect fraudulent content. The efficacy of the proposed network is validated by conducting several experiments on four real-world fake news datasets: Twitter, Jruvika Fake News Dataset, Pontes Fake News Dataset, and Risdal Fake News Dataset using multiple evaluation metrics. The results show that the proposed method outperforms the baseline methods on all four datasets. Further, the computation time of the model is also lower than the state-of-the-art methods.

cs.MM↗

arXiv2018

Polymorphism in Bi-based perovskite oxides: a first-principles study

Under normal conditions, bulk crystals of BiScO$_3$ , BiCrO$_3$, BiMnO$_3$, BiFeO$_3$, and BiCoO$_3$ present three very different variations of the perovskite structure: an antipolar phase, a rhombohedral phase with a large polarization along the space diagonal of the pseudocubic unit cell, and a supertetragonal phase with even larger polarization. With the aim of understanding the causes for this variety, we have used a genetic algorithm to search for minima in the surface energy of these materials. Our results show that the number of these minima is very large when compared to that of typical ferroelectric perovskites like BaTiO$_3$ and PbTiO$_3$ , and that a fine energy balance between them results in the large structural differences seen. As byproducts of our search we have identified charge-ordering structures with low energy in BiMnO$_3$ , and several phases with energies that are similar to that of the ground state of BiCrO$_3$. We have also found that a inverse supertetragonal phase exists in bulk, likely to be favored in films epitaxially grown at large values of tensile misfit strain.

cond-mat.mtrl-sci↗

arXiv2016

Tellurium-bridged two-leg spin ladder in Ba$_2$CuTeO$_6$

We present single-crystal growth and magnetic property studies of tellurium-bridged copper spin-1/2 system Ba$_2$CuTeO$_6$. The spin-exchange interaction among copper spins via Cu-O-Te-O-Cu super-superexchange route leads to a novel two-leg spin ladder system. Spin susceptibility $χ$(T) data indicates that the triclinic Ba$_2$CuTeO$_6$ undergoes a stepwise crossover for exchange couplings revealed by a broad maximum near T$_{max}$$\sim$75~K and an anisotropic cusp in $\frac{dχ}{dT}$(T) to signify a three dimensional (3D) antiferromagnetic long-range ordering (LRO). The 3D LRO has been suggested from the anisotropic behavior of $χ$(T) with strong c-axis spin anisotropy and the signature of spin flop transition from the isothermal magnetization below $T_N$. Analysis of magnetic heat capacity (Cm) at $T_N$~15 K indicates that most of the spin entropy (~92% ) has already been released above $T_N$, which supports the picture of consecutive spin entropy reduction upon cooling with Te-bridged two-leg spin ladder system with strong intraladder and interladder couplings. Theoretical DFT+U calculations have been performed to search for the ground state magnetic configuration and also to evaluate exchange coupling constants that support the magnetic model deduced from the combined spin susceptibility and crystal structure symmetry analysis.

cond-mat.str-el↗

arXiv2014

Peierls transition and edge reconstruction in phosphorene nanoribbons

Atomic and electronic structures of phosphorene nanoribbons are studied within density functional theory. These novel materials present different physical phenomena expected in two very different physical systems: one dimensional metallic chains and semiconductor surfaces. While `rugged' nanoribbons are semiconducting in their layer-terminnated structures, pure `linear' and `zigzag' nanoribbons are metallic due to metallic edge states. Linear nanoribbons undergo edge reconstruction and zigzag nanoribbons beyond a certain width undergo Peierls transition leading to opening of a band gap in the electronic structure and lowering of total energy. Mixed nanoribbons with linear and zigzag edges on the two sides turn out to be a curious case that display both edge reconstruction and Peierls transition simultaneously. Most phosphoeren nanoribbons turn out to be semiconductors having important implications for their application.

cond-mat.mtrl-sci↗